No abstract

[1] The study of the collapse of a granular step is of great interest for understanding transient dense granular flow conditions and for modeling geophysical flows in granular materials. We present the results of a series of finite elements simulations considering variable column aspect ratios and properties for an elastoplastic material with a Mohr-Coulomb yield rule and nonassociate flow rule. The adopted approach does not suffer limitations of typical shallow water equation methods, being able to consider strong vertical motion components. Transition from initial instability to complete flow development is simulated for columns with different aspect ratios (a 20). Simulation results are compared to original tests and available well-documented experimental data, in terms of flow development, duration, profile geometry, velocity distribution, erosion and deposition, and evolution of the interface between static and moving material. Tests involving a thick erodible layer have been performed and numerical simulation results are compared also with a real case study. Numerical results support both those of qualitative and theoretical models and the proposed general scaling laws and clarify the dependence on frictional properties. Power laws describe the normalized runout versus aspect ratio (a > 4) relationship with constants of proportionality dependent on internal friction angle and exponents ranging between 0.68 and 0.77, in good agreement with experimental results. Total duration and evolution in three successive phases agree with observations. Time for the flow front to cease motion with respect to aspect ratio is best represented by the 3.68a 0.448 relationships for a 30°internal friction angle material.

Abstract.Modelling of flow-like landslides is one of the possible approaches that can be used to simulate landslide instability and flow development. Models based on continuum mechanics and associated with a versatile rheological model are usually preferred to predict landslide runout and relevant parameters. A different approach has been used in this research. We have developed a 2-D/3-D finite element code to analyse slope stability and to model runout of mass movements characterised by very large displacements. The idea was to be able to use different material laws already known, tested and verified for granular materials. The implemented materials laws include classical elasto-plasticity, with a linear elastic part and different applicable yield surfaces with associated and non-associated flow rules. The application of Finite Element methods to model landslide run-out, contrasts previous research where typically depth-averaged equivalent-fluid approaches were adopted. The code has been applied to the simulation of large rock avalanches and rapid dry flows in different materials and under different geological and geomorphological conditions.

Landslides can occur in different environments and can interact with or fall into water reservoirs or open sea with different characteristics. The subaerial evolution and the transition from subaerial to subaqueous conditions can strongly control the landslide evolution and the generated impulse waves, and consequently the final hazard zonation. We intend to model the landslide spreading, the impact with the water surface and the generation of the impulse wave under different 2D and 3D conditions and settings. We verify the capabilities of a fully 2D and 3D FEM ALE approach to model and analyse near-field evolution. To this aim we validate the code against 2D laboratory experiments for different Froude number conditions (Fr = 1.4, 3.2). Then the Vajont rockslide (Fr = 0.26-0.75) and the consequent impulse wave are simulated in 2D and 3D. The sliding mass is simulated as an elasto-plastic Mohr-Coulomb material and the lake water as a fully inviscid low compressibility fluid. The rockslide model is validated against field observations, including the total duration, the profile and internal geometry of the final deposit, the maximum water run-up on the opposite valley flank and on the rockslide mass. 2D models are presented for both the case of a dry valley and that of the impounded lake. The set of fully 3D simulations are the first ones available and considering the rockslide evolution, propagation and interaction with the water reservoir. Advantages and disadvantages of the modelling approach are discussed.

An erodible substrate and a sharp slope break affect the dynamics and deposition of long runout landslides. We study the flow evolution of a granular mass (1.5-5.1 litres of sand or gravel) released on a bilinear chute, i.e., an incline (between 35° and 66°) followed by a horizontal sector, either sand-free or covered (1-2 cm thick sand layer). Monitoring the time evolution of the falling mass profiled at 120 Hz, the impact dynamics, erosion of the basal layer, and modes of deposition are studied. The frontal deposition is followed by a backward propagating shock wave at low slope angles (<45°), or by a forward prograding flow at greater angles. Experiments with colored sand layers show a complex sequence of dilation, folding and thrusting within both the collapsing sand flow and the substrate. Experimental results are compared with real rock avalanche data and nearly vertical collapses. The observed increase of the drop height divided by the runout (H/L or Heim's ratio) with both chute slope angle and thickness of the erodible substrate, is explained as an effect of vertical momentum loss at the slope break. Data suggest a complex evolution, different from that of a thin flow basal shear flow.To provide an approximate explanation of the dynamics, three analytical models are proposed. Erosion of a 1 cm-thick substrate is equivalent to 8-12% increase of the apparent friction coefficient. We simulate the deposition and emplacement over an erodible layer with a FEM arbitrary Lagrangian Eulerian code, and find a remarkable similarity with the time evolution observed in the experiments. 2D models evidence the internal deformation with time, 3D models simulate deposition. Capabilities and limitations of an elasto-plastic Mohr Coulomb approach for material modeling are discussed.